Questly

Turn Every Day into a Quest for Growth

Comment

Questly

Inspiration

In an age dominated by algorithm-driven content consumption and constant digital distractions, we aimed to create a solution that encourages mindful engagement. Questly was conceived as an alternative to passive scrolling — a space where users can reflect, think creatively, and learn something new, even if briefly.

What it does

Questly is a lightweight desktop application that delivers AI-generated prompts designed to encourage creative thinking, reflection, or exploration of ideas. It offers:

  • A curated list of categories selected by the user.
  • AI-generated micro-challenges relevant to the chosen category.
  • A simple input interface for user responses.
  • AI-generated evaluations and constructive feedback.
  • A related fact or insight to complement the prompt.

The application focuses on individual engagement without the distractions of profiles, progress tracking, or social interaction.

How we built it

  • Frontend: Developed using React with Vite and TypeScript.
  • Desktop Framework: Built on Electron, ensuring a cross-platform, installable experience.
  • Backend: A Node.js server handles API communication and securely stores environment variables.
  • AI Integration: Uses Together AI's togethercomputer/llama-2-7b-chat model for prompt generation, response evaluation, and feedback delivery.
  • Development Support: CREAO was used to refine and optimize certain functionalities — including improving asynchronous request handling between the Electron frontend and backend, and enhancing the logic for structured AI feedback formatting to ensure more consistent and readable responses.

The architecture prioritizes simplicity, performance, and maintainability.

Challenges we ran into

  • Ensuring secure and efficient communication between the Electron frontend and the AI backend.
  • Designing effective prompt templates to consistently yield meaningful AI-generated content.
  • Avoiding feature bloat while maintaining a high-quality user experience.
  • Handling asynchronous interactions with remote AI models while keeping latency low.

Accomplishments that we're proud of

  • Successfully integrating a large language model into a local-first desktop application.
  • Achieving fast and relevant prompt generation across diverse categories.
  • Creating a distraction-free interface focused solely on thoughtful interaction.
  • Maintaining clean architecture and efficient code throughout the stack.

What we learned

  • Clean design and minimalism are powerful but require rigorous decision-making.
  • Prompt engineering has a significant impact on AI response quality.
  • Securely working with remote LLM APIs in Electron demands careful configuration.
  • Building a focused application requires saying “no” to many otherwise useful features.

What's next for Questly

  • Enhancing prompt personalization based on user input patterns.
  • Providing optional export functionality for users to archive their responses.
  • Exploring mobile and offline versions, depending on user interest and technical feasibility.

Built With

Share this project:

Updates